EVALUATION OF NON‑INVASIVE BIOMARKERS FOR EARLY DETECTION OF COLORECTAL NEOPLASIA

Abstract

Colorectal cancer is one of the leading causes of cancer related death in the world and the earlier the diagnosis is carried out, the more the patient prognosis. The existing screening modalities do not possess the following characteristics: invasiveness, costliness and sensitivity to premalignant lesions and thus, novel non-invasive biomarkers need to be created. This study was meant to determine and validate a panel of serum-based metabolic biomarker to identify and grade colorectal neoplasia between the advanced adenoma and late-stage colorectal cancer. It was a multi-centre case-control study that included 450 people which was stratified in four groups- healthy controls, advanced adenoma , early-stage colorectal cancer  and late-stage colorectal cancer . Targeted metabolomics were done by the ultra-high-performance liquid chromatography-tandem mass spectrometry and gas chromatography-time-of-flight mass spectrometry. The choice of the biomarkers was made based on a least absolute shrinkage and selection operator (LASSO) regression model and diagnostic models have been built based on selection model and receiver operating characteristic analysis, calibration measures and compared them to conventional clinical biomarkers.These 25 metabolites were chosen and the AUC of the receiver operating characteristic curve of the following were 0.874 (advanced adenoma vs. control), 0.921 (early-stage cancer vs. control) and 0.953 (late-stage cancer vs. control). This model had a good calibration , and performed far much better on all the comparisons with carcinoembryonic antigen and carbohydrate antigen 19-9. Key discriminatory metabolites included upregulated sphingosine-1-phosphate  and acetylputrescine , and downregulated indole-3-propionic acid  and L-arginine . Pathway enrichment analysis showed that there were major perturbations in sphingolipid metabolism, the metabolism of arginine and proline as well as primary bile acid biosynthesis. The research confirms a serum metabolomic panel which has shown high diagnostic sensitivity in identifying advanced adenomas and early colorectal cancer and is better than traditional biomarkers. The results demonstrate clinical potential of the non-invasive method to enhance the early diagnosis of the disease, risk-stratification and detection of colorectal neoplasia.